You are here
The Fiorillo index -- a new index to attempt to spot citation gaming
An index has been proposed by Fiorillo (document attached).
The meaning of Fiorillo's index is simple. If h-index indicates the number of articles that are cited each at least h times, it is clear that maximum efficiency is achieved with h-index^2 citations. Fiorillo index is h_index^2/tot citation *100. So if someone has Fiorillo index =100, it means that he has maximum efficiency, an absurd limit to reach without a prior agreement between all scientists! Fiorillo found that out of 200 thousand researchers of reasonable validity taken from the Ioannidis database in Scopus of the top 2% researchers in the world, the average is around 25, and therefore he considered it suspicious (he says "BAD") that someone has an efficiency greater than 32, close to the optimal absurd one. That's all. Maybe it's his BAD index, namely we could postpone the threshold to 40 or 50, but certainly the closer to 100, the more worrying the situation is that there is citation gaming, and abuse of self-citations.
Fiorillo also makes a borderline case, having 1 article cited once also has this Fiorillo index equal to 100, but you will agree with him that it is a ridiculous case for the top 200 thousand researchers we are talking about. You can easily measure Fiorillo index in your university using the Stanford database, which you can find at
https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.30...
Regards
Mike
| Attachment | Size |
|---|---|
 FI-SCORE-FIORILLO.pdf | 743.04 KB |
 FiorilloScore-selfcitations.jpg | 156.76 KB |
| FiorilloScore-Rank-Ioannidis.jpg | 147.74 KB |
- Mike Ciavarella's blog
- Log in or register to post comments
- 4070 reads
Comments
I have added a correlation with the Stanford database ranking
See in the main post the image posted as attachment "FiorilloScore-Rank-Ioannidis.jpg" where you can see that being in the top 100 000 scientists in the world according to Ioannidis Stanford ranking is incompatible with having FiorilloScore>40 ----- hence FiorilloScore is a very quick tool to spot low quality scientists, independently on their discipline, with a calculation far simpler than the Ioannidis calculation involving a 5 dimensional parameter space. Moreover, many scientists do not enter in that database, and for them the Fiorillo score is even more appropriate.
You also find as attachment another figure, "FiorilloScore-selfcitations.jpg" where a correlation is attempted with % of self-citations. There is a clear trend to have higher Fiorillo index if you have very high % of self-citations, confirming that the Fiorello index is a good mix to spot citation gaming and also self-citation excesses...
Ioannidis from Stanford has a paper also about Fiorillo index
Ioannidis has proposed (after Fiorillo) a very similar index (the inverse of FI) as a marker of small scale orchestration, see: https://arxiv.org/abs/2406.19219 I think one has to be cautious though about interpreting values that are not extreme.
A note on Fiorillo index for not misusing it !
Dear all
having created a bit of confusion about the Fiorillo index before the summer break, with a couple of colleagues who were also sorry to be "suspected" by the Fiorillo index of having done citation gaming and orchestration, I wanted to delve deeper into the issue and apologize to my colleagues by writing a small note, in which I correct Fiorillo's conclusions and suggest not to apply this index except for statistical or very preliminary studies, to be integrated with in-depth qualitative analyses of suspected cases. I hope the reading is interesting. See https://www.researchgate.net/publication/383532645
regards, MC